05000309/LER-1997-004, Forwards Comments on Preliminary Accident Sequence Precursor Analysis of Operational Condition Discovered at Myaps on 970122 & Reported in LER 97-004

Forwards Comments on Preliminary Accident Sequence Precursor Analysis of Operational Condition Discovered at Myaps on 970122 & Reported in LER 97-004

ML20198J952
Person / Time
Site:	Maine Yankee
Issue date:	10/15/1997
From:	Hebert J Maine Yankee
To:	NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM)
References
JRH-97-213, MN-97-115, NUDOCS 9710220248
Download: ML20198J952 (4)

LER-2097-004, Forwards Comments on Preliminary Accident Sequence Precursor Analysis of Operational Condition Discovered at Myaps on 970122 & Reported in LER 97-004

Event date:
Report date:
3092097004R00 - NRC Website

text

---

6- .

MaineYankee MLM((UICM[EIN DSGII

. . 3?9 DATH ROAD

• BRUNSWICK. MAINE 04011 * (207) 79&4100 October 15,1997 MN-97115 JRI197 213 UNITED STATES NUCLEAR REGULATORY COMMISSION Attention: Document Control Desk Washington, DC 20555

References:

(a) License No. DPR 36 (Docket No. 50 309)

(b) Letter: D.ll. Dorman, USNRC to M.it Sellman, MYAPC; Review of Preliminary Accident Scluence Precursor Analysis of Operational Consideration at Maine Yankee Atomic Power Station; dated September 3,1997

Subject:

Review of Preliminary Accident Sequence Precursor Analysis of Operational Condition at Maine Yankee Atomic Power Station Gentlemen:

Reference (b) transmitted for our review and comment the preliminary Accident Sequence Precursor (ASP) analysis of an operational condition which was discovered at the Mainc Yankee Atomic Power Station on January 22,1997 and reported in Licensec Event Report (LER) No. 50 309/004.

Attachment A to this letter delineates our comments. We generally agree that the NRC model is appropriate. We do however have reservations regarding the success criteria and the assumptions which are the elements of the sequence which accounts for almost 100% of the total contribution to the increase in the Core Damage Probability.

Should you have any questions or require additional information, please contact us.

Very truly yours, at-o k. h -

James R. lichett, Manager Regulatory Affairs Department jgl3 attachment c:

Mr. Ilubert Miller Mr. D.11. Dorman Mr. R. A. Rasmussen Mr. Clough Toppan 91

'*]

Mr. P. J. Dostic giglg .[ l

$kkk,k Mr. Uldis Vanan 9710220240 97to15

. DR ADOCK-030003 9.

-- - ~ - - - - -

l Comments on Preliminary Accident Sequence Precursor (Asp) Analysis of Operational Condition Reported in LElt 50-309/97 004 Summary The " analysis results" indicate a Core Damage Probability (CDP) increase of 1.3E-05. A single

" sequence accounts for almost 100% of the total contribution to the increase in the CDP." It is also stated that "There is substantial uncertainty in this estimate because of the uncertainty in the frequency of a large I.OCA (none have occuned) and the likelihood of h10V's failing under post-1.OCA conditions."

In fact this " result" is considerably more uncertain than stated; it is based almost solely on assumptions; it probably represents a significant overestimate of Core Damage Probability, Response to " Guidance for Lictmse Review of Prdiminary ASP Analysis" The guidance provided asks three questions. A brief response to each is provided here with additional discussion following (note that question 3 has been separated into its three pads):

1. Does the " Event Description" section accurately describe the event as it occurred?

In general, this section reasonably reflects the LER's event description however two points are unclear:

i. the event involved identification of the palttitial for thennal pressure locking of certain hiOVs; the discussion makes it clear that hydraube pressure locking did not occur; it is not clear that thennal pressure locking did not occur either.

ii. most importantly, the description states that "lloron precipitation ... could lead to core damage ..."; although this generally reflects LER supplied information, it may lead to inappropriate success criteria for a realistic risk analysh; please see the " success criteria" discussion below.

2. Does the " Additional Event-Related Infonnation" section provide accurate additional information ... ?

Yes, but it may be incomplete; please see " additional model related comments" below.

3a. Does the "hiodeling Assumptions" section accurately describe the modeling done for the event?

Yes.

• a 3b. Is the modeling of the event appropriate for the events that occmred or had the potential to occur under the event conditions?

No.

The success criteria appears inappropriate; please see " success criteria" discussion below.

The analysis result is driven almost entirely by assumptions; these assumptions have little, if any, basis; no sensitivity analysis is provided to consider the effect of these assumptions on the results; please see " assumptions" below.

3c. This also includes assumptions regarding the likelihood of equipment recovery.

The analysis unrealistically assumes no credit for any recovery action; please see "acsumptions" below.

StitsnLCrikIia Maine Yankee's design basis accident analysis does require hot leg recirculation within 19 hours2.199074e-4 days <br />0.00528 hours <br />3.141534e-5 weeks <br />7.2295e-6 months <br /> of a large cold leg LOCA. The potential for this not to be provided due to thennal pressure locking of MOVs was reported via the subject LER. The success criteria involved in such design basis analysis is well known to be consenative, oRen very conservative. it is common for probabilistic risk analysis to use alternate, more realistic success criteria. For example, as stated in the subject analysis,"llot-leg recirculation is not addressed in the (Maine Yankee] IPE; however, the LER for this event indicated that the switch-over to hot leg recirculation is required 19 h aner a large cold leg break."

The subject analysis is based on the LER statement which reneets design basis. Such conservative design basis success criteria is inappropriate for use in risk analysis, especially when that analysis is stated to be "as realistic as possible" (September 3,1997 transmittal letter, D. Donnan to M.

Sellman).

While it may be possible to experience some limited, local damage to fuel in the event of boron precipitation following failure of hot leg recirculation, it may be unrealistic to as;ume signincant

" core damage" (in the sense PRAs / IPEs use the tenn). It would seem reasonable to expect, for example, that precipitated boron might go back into solution in the event of core heatup. If more realistic success criteria were considered, rather than reliance on conservative design basis criteria, this analysis might find the CDP contribution to be zero.

Funher discussion related to success criteria "2." on page 3 of the analysis can be found below under

" Additional Model Related Comments".

> e .

Assurnptiom The " analysis results" indicate a Corr, Damage Probability (CDP) increase of 1.3E-05. A single

" sequence accounts for almost 100% of the total contribution to the increase in the CDP."

In addition to the success criteria assumption discussed above, the single sequence is calculated from he following:

Sequence Element Value Source Lmge LOCA frequency 2.7E-04/yr Maine Yankee IPE Fraction of Large LOCAs in 0.5 Assumption Cold Leg Plant Availability 0.7 Not stated (but agrees with IPE value) llot Leg Recirculation Failure 0.1 Assumption Non Recovery Probability 1.0 Assumption The " analysis results" state that "There is substantial uncertainty in this estimate because of the uncertainty in the frequency of a large LOCA (none have occurred) and the likelihood of MOV's failing under post-LOCA conditions."

in fact this " result" is considerably more uncertain than s;ated; it is based almost solely on assumptions.

Additional Model Related Comments Although not significant to this particular analysis:

a. post-RAS / post-LOCA recirculation cooling, with Containment Spray pumps feeding Charging / IIPSI pumps is train dependent. That is Train A Spray feeds Train A lIPSI only (similar for B). The event tree analysis included appears to consider a "1 out of 2" success criteria for top event "llPSl" and for top event "COLDLEG". It is unclear w hether the train-wise dependence has been modeled. Successful cold leg recirculation requires the same train ofIIPSI and Spray to be operating (cross train alignment is possible with manual recovery action).

b. Success criteria "2." on page 3 of the subject analysis indicates that Large LOCA can be successfully mitigated with "no llPSI, two LPSI, one accumulator, and successful switch-over to cold leg recirculation." Since cold leg recirculation requires llPSI, it can not be accomplished with "no llPSI". Although this dependence is acknowledged later on page 3 under "HPSl",it is unclear how this is actually modeled, it would appear that sequences 7 through 9 of the event tree (Figure 2, page 7) are impossible; if so, all sequences below 6 i

could be condensed into a single sequence with IIPSI failure leading to "CD".

..-